Date: Wed, 31 Jul 2019 02:27:18 +0000 From: bugzilla-noreply@freebsd.org To: bugs@FreeBSD.org Subject: [Bug 239550] std::thread::id uses pthread_equal in undefined ways not handled by _pthread_stubs Message-ID: <bug-239550-227@https.bugs.freebsd.org/bugzilla/>
index | next in thread | raw e-mail
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=239550 Bug ID: 239550 Summary: std::thread::id uses pthread_equal in undefined ways not handled by _pthread_stubs Product: Base System Version: CURRENT Hardware: Any OS: Any Status: New Severity: Affects Many People Priority: --- Component: bin Assignee: bugs@FreeBSD.org Reporter: andrew@tao11.riddles.org.uk Created attachment 206172 --> https://bugs.freebsd.org/bugzilla/attachment.cgi?id=206172&action=edit testcase (c++) std::thread::id defines a distinguished "not equal to any actual thread" value (the implementation uses 0). But it feels entitled to pass this value to pthread_equal, whose own spec has no such distinguished value. (Per the pthread spec this gives an undefined result.) If libthr is linked in, this isn't a problem because 0 doesn't actually compare equal to any thread. But the _pthread_stubs implementation of pthread_equal returns true unconditionally, which is valid by the pthread spec (since there is no distinguished pthread_t value). The attached test program demonstrates the failure; it runs successfully if linked with -pthread, asserts out if not. The simplest fix would be to do an actual comparison in the stub version of pthread_equal. A more correct fix would be for std::thread::id not to assume it can pass its 0 value to pthread_equal, but I think this would be unnecessary overhead. -- You are receiving this mail because: You are the assignee for the bug.help
Want to link to this message? Use this
URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?bug-239550-227>